diff --git a/interface/src/Application.cpp b/interface/src/Application.cpp
index d6a90b36f8..9a7f33fda3 100644
--- a/interface/src/Application.cpp
+++ b/interface/src/Application.cpp
@@ -1176,7 +1176,6 @@ bool Application::event(QEvent* event) {
 }
 
 bool Application::eventFilter(QObject* object, QEvent* event) {
-
     if (event->type() == QEvent::ShortcutOverride) {
         if (DependencyManager::get<OffscreenUi>()->shouldSwallowShortcut(event)) {
             event->accept();
@@ -1792,6 +1791,7 @@ void Application::checkFPS() {
 }
 
 void Application::idle() {
+    PROFILE_RANGE(__FUNCTION__);
     static SimpleAverage<float> interIdleDurations;
     static uint64_t lastIdleEnd{ 0 };
 
diff --git a/interface/src/GLCanvas.cpp b/interface/src/GLCanvas.cpp
index 0eda678fa9..aff7c75cc9 100644
--- a/interface/src/GLCanvas.cpp
+++ b/interface/src/GLCanvas.cpp
@@ -58,6 +58,7 @@ void GLCanvas::initializeGL() {
 }
 
 void GLCanvas::paintGL() {
+    PROFILE_RANGE(__FUNCTION__);
     if (!_throttleRendering && !Application::getInstance()->getWindow()->isMinimized()) {
         Application::getInstance()->paintGL();
     }
diff --git a/libraries/entities-renderer/src/RenderableModelEntityItem.cpp b/libraries/entities-renderer/src/RenderableModelEntityItem.cpp
index ae1c97f07f..85b7bafc78 100644
--- a/libraries/entities-renderer/src/RenderableModelEntityItem.cpp
+++ b/libraries/entities-renderer/src/RenderableModelEntityItem.cpp
@@ -161,6 +161,7 @@ namespace render {
     template <> void payloadRender(const RenderableModelEntityItemMeta::Pointer& payload, RenderArgs* args) {
         if (args) {
             if (payload && payload->entity) {
+                PROFILE_RANGE("MetaModelRender");
                 payload->entity->render(args);
             }
         }
diff --git a/libraries/entities/src/EntityItem.cpp b/libraries/entities/src/EntityItem.cpp
index 183cbedd4f..ebe16aee71 100644
--- a/libraries/entities/src/EntityItem.cpp
+++ b/libraries/entities/src/EntityItem.cpp
@@ -12,9 +12,11 @@
 #include "EntityItem.h"
 
 #include <QtCore/QObject>
+#include <QtEndian>
 
 #include <glm/gtx/transform.hpp>
 
+#include <BufferParser.h>
 #include <ByteCountCoding.h>
 #include <GLMHelpers.h>
 #include <Octree.h>
@@ -354,50 +356,78 @@ int EntityItem::readEntityDataFromBuffer(const unsigned char* data, int bytesLef
     // ~27-35 bytes...
     const int MINIMUM_HEADER_BYTES = 27;
 
-    int bytesRead = 0;
     if (bytesLeftToRead < MINIMUM_HEADER_BYTES) {
         return 0;
     }
 
+    int clockSkew = args.sourceNode ? args.sourceNode->getClockSkewUsec() : 0;
+
+    BufferParser parser(data, bytesLeftToRead);
+
+#ifdef DEBUG
+#define VALIDATE_ENTITY_ITEM_PARSER 1
+#endif
+
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
+    int bytesRead = 0;
     int originalLength = bytesLeftToRead;
     // TODO: figure out a way to avoid the big deep copy below.
     QByteArray originalDataBuffer((const char*)data, originalLength); // big deep copy!
-
-    int clockSkew = args.sourceNode ? args.sourceNode->getClockSkewUsec() : 0;
-
     const unsigned char* dataAt = data;
+#endif
 
     // id
-    QByteArray encodedID = originalDataBuffer.mid(bytesRead, NUM_BYTES_RFC4122_UUID); // maximum possible size
-    _id = QUuid::fromRfc4122(encodedID);
-    dataAt += encodedID.size();
-    bytesRead += encodedID.size();
+    parser.readUuid(_id);
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
+    {
+        QByteArray encodedID = originalDataBuffer.mid(bytesRead, NUM_BYTES_RFC4122_UUID); // maximum possible size
+        QUuid id = QUuid::fromRfc4122(encodedID);
+        dataAt += encodedID.size();
+        bytesRead += encodedID.size();
+        Q_ASSERT(id == _id);
+        Q_ASSERT(parser.offset() == bytesRead);
+    }
+#endif
 
     // type
+    parser.readCompressedCount<quint32>((quint32&)_type);
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
     QByteArray encodedType = originalDataBuffer.mid(bytesRead); // maximum possible size
     ByteCountCoded<quint32> typeCoder = encodedType;
     encodedType = typeCoder; // determine true length
     dataAt += encodedType.size();
     bytesRead += encodedType.size();
     quint32 type = typeCoder;
-    _type = (EntityTypes::EntityType)type;
+    EntityTypes::EntityType oldType = (EntityTypes::EntityType)type;
+    Q_ASSERT(oldType == _type);
+    Q_ASSERT(parser.offset() == bytesRead);
+#endif
 
     bool overwriteLocalData = true; // assume the new content overwrites our local data
+    quint64 now = usecTimestampNow();
 
     // _created
-    quint64 createdFromBuffer = 0;
-    memcpy(&createdFromBuffer, dataAt, sizeof(createdFromBuffer));
-    dataAt += sizeof(createdFromBuffer);
-    bytesRead += sizeof(createdFromBuffer);
-
-    quint64 now = usecTimestampNow();
-    if (_created == UNKNOWN_CREATED_TIME) {
-        // we don't yet have a _created timestamp, so we accept this one
-        createdFromBuffer -= clockSkew;
-        if (createdFromBuffer > now || createdFromBuffer == UNKNOWN_CREATED_TIME) {
-            createdFromBuffer = now;
+    {
+        quint64 createdFromBuffer = 0;
+        parser.readValue(createdFromBuffer);
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
+        {
+            quint64 createdFromBuffer2 = 0;
+            memcpy(&createdFromBuffer2, dataAt, sizeof(createdFromBuffer2));
+            dataAt += sizeof(createdFromBuffer2);
+            bytesRead += sizeof(createdFromBuffer2);
+            Q_ASSERT(createdFromBuffer2 == createdFromBuffer);
+            Q_ASSERT(parser.offset() == bytesRead);
+        }
+#endif
+        if (_created == UNKNOWN_CREATED_TIME) {
+            // we don't yet have a _created timestamp, so we accept this one
+            createdFromBuffer -= clockSkew;
+            if (createdFromBuffer > now || createdFromBuffer == UNKNOWN_CREATED_TIME) {
+                createdFromBuffer = now;
+            }
+            _created = createdFromBuffer;
         }
-        _created = createdFromBuffer;
     }
 
     #ifdef WANT_DEBUG
@@ -417,15 +447,21 @@ int EntityItem::readEntityDataFromBuffer(const unsigned char* data, int bytesLef
     #endif
 
     quint64 lastEditedFromBuffer = 0;
-    quint64 lastEditedFromBufferAdjusted = 0;
 
     // TODO: we could make this encoded as a delta from _created
     // _lastEdited
-    memcpy(&lastEditedFromBuffer, dataAt, sizeof(lastEditedFromBuffer));
-    dataAt += sizeof(lastEditedFromBuffer);
-    bytesRead += sizeof(lastEditedFromBuffer);
-    lastEditedFromBufferAdjusted = lastEditedFromBuffer - clockSkew;
-
+    parser.readValue(lastEditedFromBuffer);
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
+    {
+        quint64 lastEditedFromBuffer2 = 0;
+        memcpy(&lastEditedFromBuffer2, dataAt, sizeof(lastEditedFromBuffer2));
+        dataAt += sizeof(lastEditedFromBuffer2);
+        bytesRead += sizeof(lastEditedFromBuffer2);
+        Q_ASSERT(lastEditedFromBuffer2 == lastEditedFromBuffer);
+        Q_ASSERT(parser.offset() == bytesRead);
+    }
+#endif
+    quint64 lastEditedFromBufferAdjusted = lastEditedFromBuffer - clockSkew;
     if (lastEditedFromBufferAdjusted > now) {
         lastEditedFromBufferAdjusted = now;
     }
@@ -487,9 +523,21 @@ int EntityItem::readEntityDataFromBuffer(const unsigned char* data, int bytesLef
     }
 
     // last updated is stored as ByteCountCoded delta from lastEdited
-    QByteArray encodedUpdateDelta = originalDataBuffer.mid(bytesRead); // maximum possible size
-    ByteCountCoded<quint64> updateDeltaCoder = encodedUpdateDelta;
-    quint64 updateDelta = updateDeltaCoder;
+    quint64 updateDelta;
+    parser.readCompressedCount(updateDelta);
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
+    {
+        QByteArray encodedUpdateDelta = originalDataBuffer.mid(bytesRead); // maximum possible size
+        ByteCountCoded<quint64> updateDeltaCoder = encodedUpdateDelta;
+        quint64 updateDelta2 = updateDeltaCoder;
+        Q_ASSERT(updateDelta == updateDelta2);
+        encodedUpdateDelta = updateDeltaCoder; // determine true length
+        dataAt += encodedUpdateDelta.size();
+        bytesRead += encodedUpdateDelta.size();
+        Q_ASSERT(parser.offset() == bytesRead);
+    }
+#endif
+
     if (overwriteLocalData) {
         _lastUpdated = lastEditedFromBufferAdjusted + updateDelta; // don't adjust for clock skew since we already did that
         #ifdef WANT_DEBUG
@@ -499,17 +547,25 @@ int EntityItem::readEntityDataFromBuffer(const unsigned char* data, int bytesLef
         #endif
     }
 
-    encodedUpdateDelta = updateDeltaCoder; // determine true length
-    dataAt += encodedUpdateDelta.size();
-    bytesRead += encodedUpdateDelta.size();
-
     // Newer bitstreams will have a last simulated and a last updated value
     quint64 lastSimulatedFromBufferAdjusted = now;
     if (args.bitstreamVersion >= VERSION_ENTITIES_HAS_LAST_SIMULATED_TIME) {
         // last simulated is stored as ByteCountCoded delta from lastEdited
-        QByteArray encodedSimulatedDelta = originalDataBuffer.mid(bytesRead); // maximum possible size
-        ByteCountCoded<quint64> simulatedDeltaCoder = encodedSimulatedDelta;
-        quint64 simulatedDelta = simulatedDeltaCoder;
+        quint64 simulatedDelta;
+        parser.readCompressedCount(simulatedDelta);
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
+        {
+            QByteArray encodedSimulatedDelta = originalDataBuffer.mid(bytesRead); // maximum possible size
+            ByteCountCoded<quint64> simulatedDeltaCoder = encodedSimulatedDelta;
+            quint64 simulatedDelta2 = simulatedDeltaCoder;
+            Q_ASSERT(simulatedDelta2 == simulatedDelta);
+            encodedSimulatedDelta = simulatedDeltaCoder; // determine true length
+            dataAt += encodedSimulatedDelta.size();
+            bytesRead += encodedSimulatedDelta.size();
+            Q_ASSERT(parser.offset() == bytesRead);
+        }
+#endif
+
         if (overwriteLocalData) {
             lastSimulatedFromBufferAdjusted = lastEditedFromBufferAdjusted + simulatedDelta; // don't adjust for clock skew since we already did that
             if (lastSimulatedFromBufferAdjusted > now) {
@@ -521,9 +577,6 @@ int EntityItem::readEntityDataFromBuffer(const unsigned char* data, int bytesLef
                 qCDebug(entities) << "        lastSimulatedFromBufferAdjusted:" << debugTime(lastSimulatedFromBufferAdjusted, now);
             #endif
         }
-        encodedSimulatedDelta = simulatedDeltaCoder; // determine true length
-        dataAt += encodedSimulatedDelta.size();
-        bytesRead += encodedSimulatedDelta.size();
     }
 
     #ifdef WANT_DEBUG
@@ -537,10 +590,26 @@ int EntityItem::readEntityDataFromBuffer(const unsigned char* data, int bytesLef
 
 
     // Property Flags
-    QByteArray encodedPropertyFlags = originalDataBuffer.mid(bytesRead); // maximum possible size
-    EntityPropertyFlags propertyFlags = encodedPropertyFlags;
-    dataAt += propertyFlags.getEncodedLength();
-    bytesRead += propertyFlags.getEncodedLength();
+    EntityPropertyFlags propertyFlags;
+    parser.readFlags(propertyFlags);
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
+    {
+        QByteArray encodedPropertyFlags = originalDataBuffer.mid(bytesRead); // maximum possible size
+        EntityPropertyFlags propertyFlags2 = encodedPropertyFlags;
+        dataAt += propertyFlags.getEncodedLength();
+        bytesRead += propertyFlags.getEncodedLength();
+        Q_ASSERT(propertyFlags2 == propertyFlags);
+        Q_ASSERT(parser.offset() == bytesRead);
+    }
+#endif
+
+#ifdef VALIDATE_ENTITY_ITEM_PARSER
+    Q_ASSERT(parser.data() + parser.offset() == dataAt);
+#else
+    const unsigned char* dataAt = parser.data() + parser.offset();
+    int bytesRead = parser.offset();
+#endif
+
 
     if (args.bitstreamVersion >= VERSION_ENTITIES_HAVE_SIMULATION_OWNER_AND_ACTIONS_OVER_WIRE) {
         // pack SimulationOwner and terse update properties near each other
@@ -549,6 +618,7 @@ int EntityItem::readEntityDataFromBuffer(const unsigned char* data, int bytesLef
         // even when we would otherwise ignore the rest of the packet.
 
         if (propertyFlags.getHasProperty(PROP_SIMULATION_OWNER)) {
+
             QByteArray simOwnerData;
             int bytes = OctreePacketData::unpackDataFromBytes(dataAt, simOwnerData);
             SimulationOwner newSimOwner;
diff --git a/libraries/entities/src/EntityItemID.cpp b/libraries/entities/src/EntityItemID.cpp
index 4e66f5d156..d882559ee3 100644
--- a/libraries/entities/src/EntityItemID.cpp
+++ b/libraries/entities/src/EntityItemID.cpp
@@ -11,7 +11,7 @@
 
 #include <QtCore/QObject>
 #include <QDebug>
-
+#include <BufferParser.h>
 #include <PacketHeaders.h>
 
 #include "RegisteredMetaTypes.h"
@@ -33,11 +33,8 @@ EntityItemID::EntityItemID(const QUuid& id) : QUuid(id)
 
 EntityItemID EntityItemID::readEntityItemIDFromBuffer(const unsigned char* data, int bytesLeftToRead) {
     EntityItemID result;
-
     if (bytesLeftToRead >= NUM_BYTES_RFC4122_UUID) {
-        // id
-        QByteArray encodedID((const char*)data, NUM_BYTES_RFC4122_UUID);
-        result = QUuid::fromRfc4122(encodedID);
+        BufferParser(data, bytesLeftToRead).readUuid(result);
     }
     return result;
 }
diff --git a/libraries/gpu/src/gpu/Batch.cpp b/libraries/gpu/src/gpu/Batch.cpp
index ee028e79e6..98d7b70db8 100644
--- a/libraries/gpu/src/gpu/Batch.cpp
+++ b/libraries/gpu/src/gpu/Batch.cpp
@@ -12,6 +12,17 @@
 
 #include <QDebug>
 
+#if defined(NSIGHT_FOUND)
+#include "nvToolsExt.h"
+
+ProfileRange::ProfileRange(const char *name) {
+    nvtxRangePush(name);
+}
+ProfileRange::~ProfileRange() {
+    nvtxRangePop();
+}
+#endif
+
 #define ADD_COMMAND(call) _commands.push_back(COMMAND_##call); _commandOffsets.push_back(_params.size());
 
 using namespace gpu;
diff --git a/libraries/gpu/src/gpu/Batch.h b/libraries/gpu/src/gpu/Batch.h
index 835e872b4a..5583294aba 100644
--- a/libraries/gpu/src/gpu/Batch.h
+++ b/libraries/gpu/src/gpu/Batch.h
@@ -26,17 +26,11 @@
 #include "Framebuffer.h"
 
 #if defined(NSIGHT_FOUND)
-    #include "nvToolsExt.h"
     class ProfileRange {
     public:
-        ProfileRange(const char *name) {
-            nvtxRangePush(name);
-        }
-        ~ProfileRange() {
-            nvtxRangePop();
-        }
+        ProfileRange(const char *name);
+        ~ProfileRange();
     };
-
     #define PROFILE_RANGE(name) ProfileRange profileRangeThis(name);
 #else
 #define PROFILE_RANGE(name)
diff --git a/libraries/gpu/src/gpu/Context.cpp b/libraries/gpu/src/gpu/Context.cpp
index 9cc6bb3cd7..51335f78df 100644
--- a/libraries/gpu/src/gpu/Context.cpp
+++ b/libraries/gpu/src/gpu/Context.cpp
@@ -33,9 +33,11 @@ bool Context::makeProgram(Shader& shader, const Shader::BindingSet& bindings) {
 }
 
 void Context::render(Batch& batch) {
+    PROFILE_RANGE(__FUNCTION__);
     _backend->render(batch);
 }
 
 void Context::syncCache() {
+    PROFILE_RANGE(__FUNCTION__);
     _backend->syncCache();
 }
\ No newline at end of file
diff --git a/libraries/render-utils/src/Model.cpp b/libraries/render-utils/src/Model.cpp
index 1b94c70e57..03140c4dfb 100644
--- a/libraries/render-utils/src/Model.cpp
+++ b/libraries/render-utils/src/Model.cpp
@@ -417,6 +417,7 @@ void Model::reset() {
 }
 
 bool Model::updateGeometry() {
+    PROFILE_RANGE(__FUNCTION__);
     bool needFullUpdate = false;
 
     bool needToRebuild = false;
@@ -690,6 +691,7 @@ void Model::recalculateMeshPartOffsets() {
 // entity-scripts to call.  I think it would be best to do the picking once-per-frame (in cpu, or gpu if possible)
 // and then the calls use the most recent such result. 
 void Model::recalculateMeshBoxes(bool pickAgainstTriangles) {
+    PROFILE_RANGE(__FUNCTION__);
     bool calculatedMeshTrianglesNeeded = pickAgainstTriangles && !_calculatedMeshTrianglesValid;
 
     if (!_calculatedMeshBoxesValid || calculatedMeshTrianglesNeeded || (!_calculatedMeshPartBoxesValid && pickAgainstTriangles) ) {
@@ -1281,6 +1283,7 @@ Blender::Blender(Model* model, int blendNumber, const QWeakPointer<NetworkGeomet
 }
 
 void Blender::run() {
+    PROFILE_RANGE(__FUNCTION__);
     QVector<glm::vec3> vertices, normals;
     if (!_model.isNull()) {
         int offset = 0;
@@ -1392,6 +1395,7 @@ void Model::snapToRegistrationPoint() {
 }
 
 void Model::simulate(float deltaTime, bool fullUpdate) {
+    PROFILE_RANGE(__FUNCTION__);
     fullUpdate = updateGeometry() || fullUpdate || (_scaleToFit && !_scaledToFit)
                     || (_snapModelToRegistrationPoint && !_snappedToRegistrationPoint);
                     
@@ -1829,6 +1833,7 @@ void Model::setupBatchTransform(gpu::Batch& batch, RenderArgs* args) {
 }
 
 AABox Model::getPartBounds(int meshIndex, int partIndex) {
+
     if (meshIndex < _meshStates.size()) {
         const MeshState& state = _meshStates.at(meshIndex);
         bool isSkinned = state.clusterMatrices.size() > 1;
@@ -1859,6 +1864,7 @@ AABox Model::getPartBounds(int meshIndex, int partIndex) {
 }
 
 void Model::renderPart(RenderArgs* args, int meshIndex, int partIndex, bool translucent) {
+ //   PROFILE_RANGE(__FUNCTION__);
     PerformanceTimer perfTimer("Model::renderPart");
     if (!_readyWhenAdded) {
         return; // bail asap
@@ -1933,7 +1939,9 @@ void Model::renderPart(RenderArgs* args, int meshIndex, int partIndex, bool tran
     pickPrograms(batch, mode, translucentMesh, alphaThreshold, hasLightmap, hasTangents, hasSpecular, isSkinned, wireframe,
                  args, locations);
 
-    updateVisibleJointStates();
+    {
+        updateVisibleJointStates();
+    }
 
     // if our index is ever out of range for either meshes or networkMeshes, then skip it, and set our _meshGroupsKnown
     // to false to rebuild out mesh groups.
@@ -2076,9 +2084,13 @@ void Model::renderPart(RenderArgs* args, int meshIndex, int partIndex, bool tran
         }
     }
     
-    _mutex.lock();
-    qint64 offset = _calculatedMeshPartOffset[QPair<int,int>(meshIndex, partIndex)];
-    _mutex.unlock();
+    qint64 offset;
+    {
+        // FIXME_STUTTER: We should n't have any lock here
+        _mutex.lock();
+        offset = _calculatedMeshPartOffset[QPair<int,int>(meshIndex, partIndex)];
+        _mutex.unlock();
+    }
 
     if (part.quadIndices.size() > 0) {
         batch.drawIndexed(gpu::QUADS, part.quadIndices.size(), offset);
diff --git a/libraries/render/CMakeLists.txt b/libraries/render/CMakeLists.txt
index ee99eb00b9..4d2be949e6 100644
--- a/libraries/render/CMakeLists.txt
+++ b/libraries/render/CMakeLists.txt
@@ -9,4 +9,16 @@ add_dependency_external_projects(glm)
 find_package(GLM REQUIRED)
 target_include_directories(${TARGET_NAME} PUBLIC ${GLM_INCLUDE_DIRS})
 
-link_hifi_libraries(shared gpu model)
\ No newline at end of file
+link_hifi_libraries(shared gpu model)
+
+if (WIN32)
+  if (USE_NSIGHT)
+    # try to find the Nsight package and add it to the build if we find it
+    find_package(NSIGHT)
+    if (NSIGHT_FOUND)
+      include_directories(${NSIGHT_INCLUDE_DIRS})
+      add_definitions(-DNSIGHT_FOUND)
+      target_link_libraries(${TARGET_NAME} "${NSIGHT_LIBRARIES}")
+    endif ()
+  endif()
+endif (WIN32)
\ No newline at end of file
diff --git a/libraries/render/src/render/Scene.cpp b/libraries/render/src/render/Scene.cpp
index a7145af4b5..268f2b6841 100644
--- a/libraries/render/src/render/Scene.cpp
+++ b/libraries/render/src/render/Scene.cpp
@@ -11,6 +11,7 @@
 #include "Scene.h"
 
 #include <numeric>
+#include "gpu/Batch.h"
 
 using namespace render;
 
@@ -167,6 +168,7 @@ void consolidateChangeQueue(PendingChangesQueue& queue, PendingChanges& singleBa
 }
  
 void Scene::processPendingChangesQueue() {
+    PROFILE_RANGE(__FUNCTION__);
     _changeQueueMutex.lock();
     PendingChanges consolidatedPendingChanges;
     consolidateChangeQueue(_changeQueue, consolidatedPendingChanges);
diff --git a/libraries/render/src/render/drawItemStatus.slv b/libraries/render/src/render/drawItemStatus.slv
index 9e2b4919ff..106a5684bb 100644
--- a/libraries/render/src/render/drawItemStatus.slv
+++ b/libraries/render/src/render/drawItemStatus.slv
@@ -22,10 +22,10 @@ uniform vec3 inBoundPos;
 uniform vec3 inBoundDim;
 uniform ivec4 inStatus;
 
-vec3 paintRainbow(float nv) {
-    float v = nv * 5.f;
+vec3 paintRainbow(float normalizedHue) {
+    float v = normalizedHue * 6.f;
     if (v < 0.f) {
-        return vec3(0.f, 0.f, 0.f);
+        return vec3(1.f, 0.f, 0.f);
     } else if (v < 1.f) {
         return vec3(1.f, v, 0.f);
     } else if (v < 2.f) {
@@ -36,8 +36,10 @@ vec3 paintRainbow(float nv) {
         return vec3(0.f, 1.f - (v-3.f), 1.f );
     } else if (v < 5.f) {
         return vec3((v-4.f), 0.f, 1.f );
+    } else if (v < 6.f) {
+        return vec3(1.f, 0.f, 1.f - (v-5.f));
     } else {
-        return vec3(1.f, 1.f, 1.f);
+        return vec3(1.f, 0.f, 0.f);
     }
 }
 
diff --git a/libraries/shared/src/BufferParser.h b/libraries/shared/src/BufferParser.h
new file mode 100644
index 0000000000..84bde2be31
--- /dev/null
+++ b/libraries/shared/src/BufferParser.h
@@ -0,0 +1,122 @@
+//
+//  Created by Bradley Austin Davis on 2015/07/08
+//  Copyright 2013-2015 High Fidelity, Inc.
+//
+//  Distributed under the Apache License, Version 2.0.
+//  See the accompanying file LICENSE or http://www.apache.org/licenses/LICENSE-2.0.html
+//
+
+#pragma once
+#ifndef hifi_BufferParser_h
+#define hifi_BufferParser_h
+
+#include <cstdint>
+
+#include <QUuid>
+#include <QtEndian>
+
+#include "GLMHelpers.h"
+#include "ByteCountCoding.h"
+#include "PropertyFlags.h"
+
+class BufferParser {
+public:
+    BufferParser(const uint8_t* data, size_t size, size_t offset = 0) :
+        _offset(offset), _data(data), _size(size) {
+    }
+
+    template<typename T>
+    inline void readValue(T& result) {
+        Q_ASSERT(remaining() >= sizeof(T));
+        memcpy(&result, _data + _offset, sizeof(T));
+        _offset += sizeof(T);
+    }
+
+    inline void readUuid(QUuid& result) {
+        readValue(result.data1);
+        readValue(result.data2);
+        readValue(result.data3);
+        readValue(result.data4);
+        result.data1 = qFromBigEndian<quint32>(result.data1);
+        result.data2 = qFromBigEndian<quint16>(result.data2);
+        result.data3 = qFromBigEndian<quint16>(result.data3);
+    }
+
+    template <typename T>
+    inline void readFlags(PropertyFlags<T>& result) {
+        // FIXME doing heap allocation
+        QByteArray encoded((const char*)(_data + _offset), remaining());
+        result.decode(encoded);
+        _offset += result.getEncodedLength();
+    }
+
+    template<typename T>
+    inline void readCompressedCount(T& result) {
+        // FIXME switch to a heapless implementation as soon as Brad provides it.
+        ByteCountCoded<T> codec;
+        _offset += codec.decode(reinterpret_cast<const char*>(_data + _offset), remaining());
+        result = codec.data;
+    }
+
+    inline size_t remaining() const {
+        return _size - _offset;
+    }
+
+    inline size_t offset() const {
+    	return _offset;
+    }
+
+    inline const uint8_t* data() const {
+        return _data;
+    }
+
+private:
+
+    size_t _offset{ 0 };
+    const uint8_t* const _data;
+    const size_t _size;
+};
+
+
+template<>
+inline void BufferParser::readValue<quat>(quat& result) {
+    size_t advance = unpackOrientationQuatFromBytes(_data + _offset, result);
+    _offset += advance;
+}
+
+template<>
+inline void BufferParser::readValue(QString& result) {
+    uint16_t length; readValue(length);
+    result = QString((const char*)_data + _offset);
+}
+
+template<>
+inline void BufferParser::readValue(QUuid& result) {
+    uint16_t length; readValue(length);
+    Q_ASSERT(16 == length);
+    readUuid(result);
+}
+
+template<>
+inline void BufferParser::readValue(xColor& result) {
+    readValue(result.red);
+    readValue(result.blue);
+    readValue(result.green);
+}
+
+template<>
+inline void BufferParser::readValue(QVector<glm::vec3>& result) {
+    uint16_t length; readValue(length);
+    result.resize(length);
+    memcpy(result.data(), _data + _offset, sizeof(glm::vec3) * length);
+    _offset += sizeof(glm::vec3) * length;
+}
+
+template<>
+inline void BufferParser::readValue(QByteArray& result) {
+    uint16_t length; readValue(length);
+    result = QByteArray((char*)_data + _offset, (int)length);
+    _offset += length;
+}
+
+#endif